AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 1 | // SPDX-License-Identifier: GPL-2.0+ and MIT |
| 2 | /* |
| 3 | * RSA library - generate parameters for a public key |
| 4 | * |
| 5 | * Copyright (c) 2019 Linaro Limited |
| 6 | * Author: AKASHI Takahiro |
| 7 | * |
| 8 | * Big number routines in this file come from BearSSL: |
| 9 | * Copyright (c) 2016 Thomas Pornin <pornin@bolet.org> |
| 10 | */ |
| 11 | |
| 12 | #include <common.h> |
| 13 | #include <image.h> |
| 14 | #include <malloc.h> |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 15 | #include <crypto/internal/rsa.h> |
| 16 | #include <u-boot/rsa-mod-exp.h> |
Robert Reither | 8479333 | 2020-09-14 13:12:02 +0200 | [diff] [blame] | 17 | #include <asm/unaligned.h> |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 18 | |
| 19 | /** |
| 20 | * br_dec16be() - Convert 16-bit big-endian integer to native |
| 21 | * @src: Pointer to data |
| 22 | * Return: Native-endian integer |
| 23 | */ |
| 24 | static unsigned br_dec16be(const void *src) |
| 25 | { |
Robert Reither | 8479333 | 2020-09-14 13:12:02 +0200 | [diff] [blame] | 26 | return get_unaligned_be16(src); |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 27 | } |
| 28 | |
| 29 | /** |
| 30 | * br_dec32be() - Convert 32-bit big-endian integer to native |
| 31 | * @src: Pointer to data |
| 32 | * Return: Native-endian integer |
| 33 | */ |
| 34 | static uint32_t br_dec32be(const void *src) |
| 35 | { |
Robert Reither | 8479333 | 2020-09-14 13:12:02 +0200 | [diff] [blame] | 36 | return get_unaligned_be32(src); |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 37 | } |
| 38 | |
| 39 | /** |
| 40 | * br_enc32be() - Convert native 32-bit integer to big-endian |
| 41 | * @dst: Pointer to buffer to store big-endian integer in |
| 42 | * @x: Native 32-bit integer |
| 43 | */ |
| 44 | static void br_enc32be(void *dst, uint32_t x) |
| 45 | { |
| 46 | __be32 tmp; |
| 47 | |
| 48 | tmp = cpu_to_be32(x); |
| 49 | memcpy(dst, &tmp, sizeof(tmp)); |
| 50 | } |
| 51 | |
| 52 | /* from BearSSL's src/inner.h */ |
| 53 | |
| 54 | /* |
| 55 | * Negate a boolean. |
| 56 | */ |
| 57 | static uint32_t NOT(uint32_t ctl) |
| 58 | { |
| 59 | return ctl ^ 1; |
| 60 | } |
| 61 | |
| 62 | /* |
| 63 | * Multiplexer: returns x if ctl == 1, y if ctl == 0. |
| 64 | */ |
| 65 | static uint32_t MUX(uint32_t ctl, uint32_t x, uint32_t y) |
| 66 | { |
| 67 | return y ^ (-ctl & (x ^ y)); |
| 68 | } |
| 69 | |
| 70 | /* |
| 71 | * Equality check: returns 1 if x == y, 0 otherwise. |
| 72 | */ |
| 73 | static uint32_t EQ(uint32_t x, uint32_t y) |
| 74 | { |
| 75 | uint32_t q; |
| 76 | |
| 77 | q = x ^ y; |
| 78 | return NOT((q | -q) >> 31); |
| 79 | } |
| 80 | |
| 81 | /* |
| 82 | * Inequality check: returns 1 if x != y, 0 otherwise. |
| 83 | */ |
| 84 | static uint32_t NEQ(uint32_t x, uint32_t y) |
| 85 | { |
| 86 | uint32_t q; |
| 87 | |
| 88 | q = x ^ y; |
| 89 | return (q | -q) >> 31; |
| 90 | } |
| 91 | |
| 92 | /* |
| 93 | * Comparison: returns 1 if x > y, 0 otherwise. |
| 94 | */ |
| 95 | static uint32_t GT(uint32_t x, uint32_t y) |
| 96 | { |
| 97 | /* |
| 98 | * If both x < 2^31 and y < 2^31, then y-x will have its high |
| 99 | * bit set if x > y, cleared otherwise. |
| 100 | * |
| 101 | * If either x >= 2^31 or y >= 2^31 (but not both), then the |
| 102 | * result is the high bit of x. |
| 103 | * |
| 104 | * If both x >= 2^31 and y >= 2^31, then we can virtually |
| 105 | * subtract 2^31 from both, and we are back to the first case. |
| 106 | * Since (y-2^31)-(x-2^31) = y-x, the subtraction is already |
| 107 | * fine. |
| 108 | */ |
| 109 | uint32_t z; |
| 110 | |
| 111 | z = y - x; |
| 112 | return (z ^ ((x ^ y) & (x ^ z))) >> 31; |
| 113 | } |
| 114 | |
| 115 | /* |
| 116 | * Compute the bit length of a 32-bit integer. Returned value is between 0 |
| 117 | * and 32 (inclusive). |
| 118 | */ |
| 119 | static uint32_t BIT_LENGTH(uint32_t x) |
| 120 | { |
| 121 | uint32_t k, c; |
| 122 | |
| 123 | k = NEQ(x, 0); |
| 124 | c = GT(x, 0xFFFF); x = MUX(c, x >> 16, x); k += c << 4; |
| 125 | c = GT(x, 0x00FF); x = MUX(c, x >> 8, x); k += c << 3; |
| 126 | c = GT(x, 0x000F); x = MUX(c, x >> 4, x); k += c << 2; |
| 127 | c = GT(x, 0x0003); x = MUX(c, x >> 2, x); k += c << 1; |
| 128 | k += GT(x, 0x0001); |
| 129 | return k; |
| 130 | } |
| 131 | |
| 132 | #define GE(x, y) NOT(GT(y, x)) |
| 133 | #define LT(x, y) GT(y, x) |
| 134 | #define MUL(x, y) ((uint64_t)(x) * (uint64_t)(y)) |
| 135 | |
| 136 | /* |
| 137 | * Integers 'i32' |
| 138 | * -------------- |
| 139 | * |
| 140 | * The 'i32' functions implement computations on big integers using |
| 141 | * an internal representation as an array of 32-bit integers. For |
| 142 | * an array x[]: |
| 143 | * -- x[0] contains the "announced bit length" of the integer |
| 144 | * -- x[1], x[2]... contain the value in little-endian order (x[1] |
| 145 | * contains the least significant 32 bits) |
| 146 | * |
| 147 | * Multiplications rely on the elementary 32x32->64 multiplication. |
| 148 | * |
| 149 | * The announced bit length specifies the number of bits that are |
| 150 | * significant in the subsequent 32-bit words. Unused bits in the |
| 151 | * last (most significant) word are set to 0; subsequent words are |
| 152 | * uninitialized and need not exist at all. |
| 153 | * |
| 154 | * The execution time and memory access patterns of all computations |
| 155 | * depend on the announced bit length, but not on the actual word |
| 156 | * values. For modular integers, the announced bit length of any integer |
| 157 | * modulo n is equal to the actual bit length of n; thus, computations |
| 158 | * on modular integers are "constant-time" (only the modulus length may |
| 159 | * leak). |
| 160 | */ |
| 161 | |
| 162 | /* |
| 163 | * Extract one word from an integer. The offset is counted in bits. |
| 164 | * The word MUST entirely fit within the word elements corresponding |
| 165 | * to the announced bit length of a[]. |
| 166 | */ |
| 167 | static uint32_t br_i32_word(const uint32_t *a, uint32_t off) |
| 168 | { |
| 169 | size_t u; |
| 170 | unsigned j; |
| 171 | |
| 172 | u = (size_t)(off >> 5) + 1; |
| 173 | j = (unsigned)off & 31; |
| 174 | if (j == 0) { |
| 175 | return a[u]; |
| 176 | } else { |
| 177 | return (a[u] >> j) | (a[u + 1] << (32 - j)); |
| 178 | } |
| 179 | } |
| 180 | |
| 181 | /* from BearSSL's src/int/i32_bitlen.c */ |
| 182 | |
| 183 | /* |
| 184 | * Compute the actual bit length of an integer. The argument x should |
| 185 | * point to the first (least significant) value word of the integer. |
| 186 | * The len 'xlen' contains the number of 32-bit words to access. |
| 187 | * |
| 188 | * CT: value or length of x does not leak. |
| 189 | */ |
| 190 | static uint32_t br_i32_bit_length(uint32_t *x, size_t xlen) |
| 191 | { |
| 192 | uint32_t tw, twk; |
| 193 | |
| 194 | tw = 0; |
| 195 | twk = 0; |
| 196 | while (xlen -- > 0) { |
| 197 | uint32_t w, c; |
| 198 | |
| 199 | c = EQ(tw, 0); |
| 200 | w = x[xlen]; |
| 201 | tw = MUX(c, w, tw); |
| 202 | twk = MUX(c, (uint32_t)xlen, twk); |
| 203 | } |
| 204 | return (twk << 5) + BIT_LENGTH(tw); |
| 205 | } |
| 206 | |
| 207 | /* from BearSSL's src/int/i32_decode.c */ |
| 208 | |
| 209 | /* |
| 210 | * Decode an integer from its big-endian unsigned representation. The |
| 211 | * "true" bit length of the integer is computed, but all words of x[] |
| 212 | * corresponding to the full 'len' bytes of the source are set. |
| 213 | * |
| 214 | * CT: value or length of x does not leak. |
| 215 | */ |
| 216 | static void br_i32_decode(uint32_t *x, const void *src, size_t len) |
| 217 | { |
| 218 | const unsigned char *buf; |
| 219 | size_t u, v; |
| 220 | |
| 221 | buf = src; |
| 222 | u = len; |
| 223 | v = 1; |
| 224 | for (;;) { |
| 225 | if (u < 4) { |
| 226 | uint32_t w; |
| 227 | |
| 228 | if (u < 2) { |
| 229 | if (u == 0) { |
| 230 | break; |
| 231 | } else { |
| 232 | w = buf[0]; |
| 233 | } |
| 234 | } else { |
| 235 | if (u == 2) { |
| 236 | w = br_dec16be(buf); |
| 237 | } else { |
| 238 | w = ((uint32_t)buf[0] << 16) |
| 239 | | br_dec16be(buf + 1); |
| 240 | } |
| 241 | } |
| 242 | x[v ++] = w; |
| 243 | break; |
| 244 | } else { |
| 245 | u -= 4; |
| 246 | x[v ++] = br_dec32be(buf + u); |
| 247 | } |
| 248 | } |
| 249 | x[0] = br_i32_bit_length(x + 1, v - 1); |
| 250 | } |
| 251 | |
| 252 | /* from BearSSL's src/int/i32_encode.c */ |
| 253 | |
| 254 | /* |
| 255 | * Encode an integer into its big-endian unsigned representation. The |
| 256 | * output length in bytes is provided (parameter 'len'); if the length |
| 257 | * is too short then the integer is appropriately truncated; if it is |
| 258 | * too long then the extra bytes are set to 0. |
| 259 | */ |
| 260 | static void br_i32_encode(void *dst, size_t len, const uint32_t *x) |
| 261 | { |
| 262 | unsigned char *buf; |
| 263 | size_t k; |
| 264 | |
| 265 | buf = dst; |
| 266 | |
| 267 | /* |
| 268 | * Compute the announced size of x in bytes; extra bytes are |
| 269 | * filled with zeros. |
| 270 | */ |
| 271 | k = (x[0] + 7) >> 3; |
| 272 | while (len > k) { |
| 273 | *buf ++ = 0; |
| 274 | len --; |
| 275 | } |
| 276 | |
| 277 | /* |
| 278 | * Now we use k as index within x[]. That index starts at 1; |
| 279 | * we initialize it to the topmost complete word, and process |
| 280 | * any remaining incomplete word. |
| 281 | */ |
| 282 | k = (len + 3) >> 2; |
| 283 | switch (len & 3) { |
| 284 | case 3: |
| 285 | *buf ++ = x[k] >> 16; |
| 286 | /* fall through */ |
| 287 | case 2: |
| 288 | *buf ++ = x[k] >> 8; |
| 289 | /* fall through */ |
| 290 | case 1: |
| 291 | *buf ++ = x[k]; |
| 292 | k --; |
| 293 | } |
| 294 | |
| 295 | /* |
| 296 | * Encode all complete words. |
| 297 | */ |
| 298 | while (k > 0) { |
| 299 | br_enc32be(buf, x[k]); |
| 300 | k --; |
| 301 | buf += 4; |
| 302 | } |
| 303 | } |
| 304 | |
| 305 | /* from BearSSL's src/int/i32_ninv32.c */ |
| 306 | |
| 307 | /* |
| 308 | * Compute -(1/x) mod 2^32. If x is even, then this function returns 0. |
| 309 | */ |
| 310 | static uint32_t br_i32_ninv32(uint32_t x) |
| 311 | { |
| 312 | uint32_t y; |
| 313 | |
| 314 | y = 2 - x; |
| 315 | y *= 2 - y * x; |
| 316 | y *= 2 - y * x; |
| 317 | y *= 2 - y * x; |
| 318 | y *= 2 - y * x; |
| 319 | return MUX(x & 1, -y, 0); |
| 320 | } |
| 321 | |
| 322 | /* from BearSSL's src/int/i32_add.c */ |
| 323 | |
| 324 | /* |
| 325 | * Add b[] to a[] and return the carry (0 or 1). If ctl is 0, then a[] |
| 326 | * is unmodified, but the carry is still computed and returned. The |
| 327 | * arrays a[] and b[] MUST have the same announced bit length. |
| 328 | * |
| 329 | * a[] and b[] MAY be the same array, but partial overlap is not allowed. |
| 330 | */ |
| 331 | static uint32_t br_i32_add(uint32_t *a, const uint32_t *b, uint32_t ctl) |
| 332 | { |
| 333 | uint32_t cc; |
| 334 | size_t u, m; |
| 335 | |
| 336 | cc = 0; |
| 337 | m = (a[0] + 63) >> 5; |
| 338 | for (u = 1; u < m; u ++) { |
| 339 | uint32_t aw, bw, naw; |
| 340 | |
| 341 | aw = a[u]; |
| 342 | bw = b[u]; |
| 343 | naw = aw + bw + cc; |
| 344 | |
| 345 | /* |
| 346 | * Carry is 1 if naw < aw. Carry is also 1 if naw == aw |
| 347 | * AND the carry was already 1. |
| 348 | */ |
| 349 | cc = (cc & EQ(naw, aw)) | LT(naw, aw); |
| 350 | a[u] = MUX(ctl, naw, aw); |
| 351 | } |
| 352 | return cc; |
| 353 | } |
| 354 | |
| 355 | /* from BearSSL's src/int/i32_sub.c */ |
| 356 | |
| 357 | /* |
| 358 | * Subtract b[] from a[] and return the carry (0 or 1). If ctl is 0, |
| 359 | * then a[] is unmodified, but the carry is still computed and returned. |
| 360 | * The arrays a[] and b[] MUST have the same announced bit length. |
| 361 | * |
| 362 | * a[] and b[] MAY be the same array, but partial overlap is not allowed. |
| 363 | */ |
| 364 | static uint32_t br_i32_sub(uint32_t *a, const uint32_t *b, uint32_t ctl) |
| 365 | { |
| 366 | uint32_t cc; |
| 367 | size_t u, m; |
| 368 | |
| 369 | cc = 0; |
| 370 | m = (a[0] + 63) >> 5; |
| 371 | for (u = 1; u < m; u ++) { |
| 372 | uint32_t aw, bw, naw; |
| 373 | |
| 374 | aw = a[u]; |
| 375 | bw = b[u]; |
| 376 | naw = aw - bw - cc; |
| 377 | |
| 378 | /* |
| 379 | * Carry is 1 if naw > aw. Carry is 1 also if naw == aw |
| 380 | * AND the carry was already 1. |
| 381 | */ |
| 382 | cc = (cc & EQ(naw, aw)) | GT(naw, aw); |
| 383 | a[u] = MUX(ctl, naw, aw); |
| 384 | } |
| 385 | return cc; |
| 386 | } |
| 387 | |
| 388 | /* from BearSSL's src/int/i32_div32.c */ |
| 389 | |
| 390 | /* |
| 391 | * Constant-time division. The dividend hi:lo is divided by the |
| 392 | * divisor d; the quotient is returned and the remainder is written |
| 393 | * in *r. If hi == d, then the quotient does not fit on 32 bits; |
| 394 | * returned value is thus truncated. If hi > d, returned values are |
| 395 | * indeterminate. |
| 396 | */ |
| 397 | static uint32_t br_divrem(uint32_t hi, uint32_t lo, uint32_t d, uint32_t *r) |
| 398 | { |
| 399 | /* TODO: optimize this */ |
| 400 | uint32_t q; |
| 401 | uint32_t ch, cf; |
| 402 | int k; |
| 403 | |
| 404 | q = 0; |
| 405 | ch = EQ(hi, d); |
| 406 | hi = MUX(ch, 0, hi); |
| 407 | for (k = 31; k > 0; k --) { |
| 408 | int j; |
| 409 | uint32_t w, ctl, hi2, lo2; |
| 410 | |
| 411 | j = 32 - k; |
| 412 | w = (hi << j) | (lo >> k); |
| 413 | ctl = GE(w, d) | (hi >> k); |
| 414 | hi2 = (w - d) >> j; |
| 415 | lo2 = lo - (d << k); |
| 416 | hi = MUX(ctl, hi2, hi); |
| 417 | lo = MUX(ctl, lo2, lo); |
| 418 | q |= ctl << k; |
| 419 | } |
| 420 | cf = GE(lo, d) | hi; |
| 421 | q |= cf; |
| 422 | *r = MUX(cf, lo - d, lo); |
| 423 | return q; |
| 424 | } |
| 425 | |
| 426 | /* |
| 427 | * Wrapper for br_divrem(); the remainder is returned, and the quotient |
| 428 | * is discarded. |
| 429 | */ |
| 430 | static uint32_t br_rem(uint32_t hi, uint32_t lo, uint32_t d) |
| 431 | { |
| 432 | uint32_t r; |
| 433 | |
| 434 | br_divrem(hi, lo, d, &r); |
| 435 | return r; |
| 436 | } |
| 437 | |
| 438 | /* |
| 439 | * Wrapper for br_divrem(); the quotient is returned, and the remainder |
| 440 | * is discarded. |
| 441 | */ |
| 442 | static uint32_t br_div(uint32_t hi, uint32_t lo, uint32_t d) |
| 443 | { |
| 444 | uint32_t r; |
| 445 | |
| 446 | return br_divrem(hi, lo, d, &r); |
| 447 | } |
| 448 | |
| 449 | /* from BearSSL's src/int/i32_muladd.c */ |
| 450 | |
| 451 | /* |
| 452 | * Multiply x[] by 2^32 and then add integer z, modulo m[]. This |
| 453 | * function assumes that x[] and m[] have the same announced bit |
| 454 | * length, and the announced bit length of m[] matches its true |
| 455 | * bit length. |
| 456 | * |
| 457 | * x[] and m[] MUST be distinct arrays. |
| 458 | * |
| 459 | * CT: only the common announced bit length of x and m leaks, not |
| 460 | * the values of x, z or m. |
| 461 | */ |
| 462 | static void br_i32_muladd_small(uint32_t *x, uint32_t z, const uint32_t *m) |
| 463 | { |
| 464 | uint32_t m_bitlen; |
| 465 | size_t u, mlen; |
| 466 | uint32_t a0, a1, b0, hi, g, q, tb; |
| 467 | uint32_t chf, clow, under, over; |
| 468 | uint64_t cc; |
| 469 | |
| 470 | /* |
| 471 | * We can test on the modulus bit length since we accept to |
| 472 | * leak that length. |
| 473 | */ |
| 474 | m_bitlen = m[0]; |
| 475 | if (m_bitlen == 0) { |
| 476 | return; |
| 477 | } |
| 478 | if (m_bitlen <= 32) { |
| 479 | x[1] = br_rem(x[1], z, m[1]); |
| 480 | return; |
| 481 | } |
| 482 | mlen = (m_bitlen + 31) >> 5; |
| 483 | |
| 484 | /* |
| 485 | * Principle: we estimate the quotient (x*2^32+z)/m by |
| 486 | * doing a 64/32 division with the high words. |
| 487 | * |
| 488 | * Let: |
| 489 | * w = 2^32 |
| 490 | * a = (w*a0 + a1) * w^N + a2 |
| 491 | * b = b0 * w^N + b2 |
| 492 | * such that: |
| 493 | * 0 <= a0 < w |
| 494 | * 0 <= a1 < w |
| 495 | * 0 <= a2 < w^N |
| 496 | * w/2 <= b0 < w |
| 497 | * 0 <= b2 < w^N |
| 498 | * a < w*b |
| 499 | * I.e. the two top words of a are a0:a1, the top word of b is |
| 500 | * b0, we ensured that b0 is "full" (high bit set), and a is |
| 501 | * such that the quotient q = a/b fits on one word (0 <= q < w). |
| 502 | * |
| 503 | * If a = b*q + r (with 0 <= r < q), we can estimate q by |
| 504 | * doing an Euclidean division on the top words: |
| 505 | * a0*w+a1 = b0*u + v (with 0 <= v < w) |
| 506 | * Then the following holds: |
| 507 | * 0 <= u <= w |
| 508 | * u-2 <= q <= u |
| 509 | */ |
| 510 | a0 = br_i32_word(x, m_bitlen - 32); |
| 511 | hi = x[mlen]; |
| 512 | memmove(x + 2, x + 1, (mlen - 1) * sizeof *x); |
| 513 | x[1] = z; |
| 514 | a1 = br_i32_word(x, m_bitlen - 32); |
| 515 | b0 = br_i32_word(m, m_bitlen - 32); |
| 516 | |
| 517 | /* |
| 518 | * We estimate a divisor q. If the quotient returned by br_div() |
| 519 | * is g: |
| 520 | * -- If a0 == b0 then g == 0; we want q = 0xFFFFFFFF. |
| 521 | * -- Otherwise: |
| 522 | * -- if g == 0 then we set q = 0; |
| 523 | * -- otherwise, we set q = g - 1. |
| 524 | * The properties described above then ensure that the true |
| 525 | * quotient is q-1, q or q+1. |
| 526 | */ |
| 527 | g = br_div(a0, a1, b0); |
| 528 | q = MUX(EQ(a0, b0), 0xFFFFFFFF, MUX(EQ(g, 0), 0, g - 1)); |
| 529 | |
| 530 | /* |
| 531 | * We subtract q*m from x (with the extra high word of value 'hi'). |
| 532 | * Since q may be off by 1 (in either direction), we may have to |
| 533 | * add or subtract m afterwards. |
| 534 | * |
| 535 | * The 'tb' flag will be true (1) at the end of the loop if the |
| 536 | * result is greater than or equal to the modulus (not counting |
| 537 | * 'hi' or the carry). |
| 538 | */ |
| 539 | cc = 0; |
| 540 | tb = 1; |
| 541 | for (u = 1; u <= mlen; u ++) { |
| 542 | uint32_t mw, zw, xw, nxw; |
| 543 | uint64_t zl; |
| 544 | |
| 545 | mw = m[u]; |
| 546 | zl = MUL(mw, q) + cc; |
| 547 | cc = (uint32_t)(zl >> 32); |
| 548 | zw = (uint32_t)zl; |
| 549 | xw = x[u]; |
| 550 | nxw = xw - zw; |
| 551 | cc += (uint64_t)GT(nxw, xw); |
| 552 | x[u] = nxw; |
| 553 | tb = MUX(EQ(nxw, mw), tb, GT(nxw, mw)); |
| 554 | } |
| 555 | |
| 556 | /* |
| 557 | * If we underestimated q, then either cc < hi (one extra bit |
| 558 | * beyond the top array word), or cc == hi and tb is true (no |
| 559 | * extra bit, but the result is not lower than the modulus). In |
| 560 | * these cases we must subtract m once. |
| 561 | * |
| 562 | * Otherwise, we may have overestimated, which will show as |
| 563 | * cc > hi (thus a negative result). Correction is adding m once. |
| 564 | */ |
| 565 | chf = (uint32_t)(cc >> 32); |
| 566 | clow = (uint32_t)cc; |
| 567 | over = chf | GT(clow, hi); |
| 568 | under = ~over & (tb | (~chf & LT(clow, hi))); |
| 569 | br_i32_add(x, m, over); |
| 570 | br_i32_sub(x, m, under); |
| 571 | } |
| 572 | |
| 573 | /* from BearSSL's src/int/i32_reduce.c */ |
| 574 | |
| 575 | /* |
| 576 | * Reduce an integer (a[]) modulo another (m[]). The result is written |
| 577 | * in x[] and its announced bit length is set to be equal to that of m[]. |
| 578 | * |
| 579 | * x[] MUST be distinct from a[] and m[]. |
| 580 | * |
| 581 | * CT: only announced bit lengths leak, not values of x, a or m. |
| 582 | */ |
| 583 | static void br_i32_reduce(uint32_t *x, const uint32_t *a, const uint32_t *m) |
| 584 | { |
| 585 | uint32_t m_bitlen, a_bitlen; |
| 586 | size_t mlen, alen, u; |
| 587 | |
| 588 | m_bitlen = m[0]; |
| 589 | mlen = (m_bitlen + 31) >> 5; |
| 590 | |
| 591 | x[0] = m_bitlen; |
| 592 | if (m_bitlen == 0) { |
| 593 | return; |
| 594 | } |
| 595 | |
| 596 | /* |
| 597 | * If the source is shorter, then simply copy all words from a[] |
| 598 | * and zero out the upper words. |
| 599 | */ |
| 600 | a_bitlen = a[0]; |
| 601 | alen = (a_bitlen + 31) >> 5; |
| 602 | if (a_bitlen < m_bitlen) { |
| 603 | memcpy(x + 1, a + 1, alen * sizeof *a); |
| 604 | for (u = alen; u < mlen; u ++) { |
| 605 | x[u + 1] = 0; |
| 606 | } |
| 607 | return; |
| 608 | } |
| 609 | |
| 610 | /* |
| 611 | * The source length is at least equal to that of the modulus. |
| 612 | * We must thus copy N-1 words, and input the remaining words |
| 613 | * one by one. |
| 614 | */ |
| 615 | memcpy(x + 1, a + 2 + (alen - mlen), (mlen - 1) * sizeof *a); |
| 616 | x[mlen] = 0; |
| 617 | for (u = 1 + alen - mlen; u > 0; u --) { |
| 618 | br_i32_muladd_small(x, a[u], m); |
| 619 | } |
| 620 | } |
| 621 | |
| 622 | /** |
| 623 | * rsa_free_key_prop() - Free key properties |
| 624 | * @prop: Pointer to struct key_prop |
| 625 | * |
| 626 | * This function frees all the memories allocated by rsa_gen_key_prop(). |
| 627 | */ |
| 628 | void rsa_free_key_prop(struct key_prop *prop) |
| 629 | { |
| 630 | if (!prop) |
| 631 | return; |
| 632 | |
| 633 | free((void *)prop->modulus); |
| 634 | free((void *)prop->public_exponent); |
| 635 | free((void *)prop->rr); |
| 636 | |
| 637 | free(prop); |
| 638 | } |
| 639 | |
| 640 | /** |
| 641 | * rsa_gen_key_prop() - Generate key properties of RSA public key |
| 642 | * @key: Specifies key data in DER format |
| 643 | * @keylen: Length of @key |
| 644 | * @prop: Generated key property |
| 645 | * |
| 646 | * This function takes a blob of encoded RSA public key data in DER |
| 647 | * format, parse it and generate all the relevant properties |
| 648 | * in key_prop structure. |
| 649 | * Return a pointer to struct key_prop in @prop on success. |
| 650 | * |
| 651 | * Return: 0 on success, negative on error |
| 652 | */ |
| 653 | int rsa_gen_key_prop(const void *key, uint32_t keylen, struct key_prop **prop) |
| 654 | { |
| 655 | struct rsa_key rsa_key; |
| 656 | uint32_t *n = NULL, *rr = NULL, *rrtmp = NULL; |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 657 | int rlen, i, ret = 0; |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 658 | |
| 659 | *prop = calloc(sizeof(**prop), 1); |
Heiko Stuebner | 49d0ea3 | 2020-07-07 22:57:26 +0200 | [diff] [blame] | 660 | if (!(*prop)) { |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 661 | ret = -ENOMEM; |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 662 | goto out; |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 663 | } |
| 664 | |
| 665 | ret = rsa_parse_pub_key(&rsa_key, key, keylen); |
| 666 | if (ret) |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 667 | goto out; |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 668 | |
| 669 | /* modulus */ |
| 670 | /* removing leading 0's */ |
| 671 | for (i = 0; i < rsa_key.n_sz && !rsa_key.n[i]; i++) |
| 672 | ; |
| 673 | (*prop)->num_bits = (rsa_key.n_sz - i) * 8; |
| 674 | (*prop)->modulus = malloc(rsa_key.n_sz - i); |
| 675 | if (!(*prop)->modulus) { |
| 676 | ret = -ENOMEM; |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 677 | goto out; |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 678 | } |
| 679 | memcpy((void *)(*prop)->modulus, &rsa_key.n[i], rsa_key.n_sz - i); |
| 680 | |
Heiko Stuebner | 49d0ea3 | 2020-07-07 22:57:26 +0200 | [diff] [blame] | 681 | n = calloc(sizeof(uint32_t), 1 + ((*prop)->num_bits >> 5)); |
| 682 | rr = calloc(sizeof(uint32_t), 1 + (((*prop)->num_bits * 2) >> 5)); |
| 683 | rrtmp = calloc(sizeof(uint32_t), 2 + (((*prop)->num_bits * 2) >> 5)); |
| 684 | if (!n || !rr || !rrtmp) { |
| 685 | ret = -ENOMEM; |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 686 | goto out; |
Heiko Stuebner | 49d0ea3 | 2020-07-07 22:57:26 +0200 | [diff] [blame] | 687 | } |
| 688 | |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 689 | /* exponent */ |
| 690 | (*prop)->public_exponent = calloc(1, sizeof(uint64_t)); |
| 691 | if (!(*prop)->public_exponent) { |
| 692 | ret = -ENOMEM; |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 693 | goto out; |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 694 | } |
| 695 | memcpy((void *)(*prop)->public_exponent + sizeof(uint64_t) |
| 696 | - rsa_key.e_sz, |
| 697 | rsa_key.e, rsa_key.e_sz); |
Heiko Stuebner | 8434b43 | 2020-06-18 16:23:23 +0200 | [diff] [blame] | 698 | (*prop)->exp_len = sizeof(uint64_t); |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 699 | |
| 700 | /* n0 inverse */ |
| 701 | br_i32_decode(n, &rsa_key.n[i], rsa_key.n_sz - i); |
| 702 | (*prop)->n0inv = br_i32_ninv32(n[1]); |
| 703 | |
| 704 | /* R^2 mod n; R = 2^(num_bits) */ |
| 705 | rlen = (*prop)->num_bits * 2; /* #bits of R^2 = (2^num_bits)^2 */ |
| 706 | rr[0] = 0; |
| 707 | *(uint8_t *)&rr[0] = (1 << (rlen % 8)); |
| 708 | for (i = 1; i < (((rlen + 31) >> 5) + 1); i++) |
| 709 | rr[i] = 0; |
| 710 | br_i32_decode(rrtmp, rr, ((rlen + 7) >> 3) + 1); |
| 711 | br_i32_reduce(rr, rrtmp, n); |
| 712 | |
| 713 | rlen = ((*prop)->num_bits + 7) >> 3; /* #bytes of R^2 mod n */ |
| 714 | (*prop)->rr = malloc(rlen); |
| 715 | if (!(*prop)->rr) { |
| 716 | ret = -ENOMEM; |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 717 | goto out; |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 718 | } |
| 719 | br_i32_encode((void *)(*prop)->rr, rlen, rr); |
| 720 | |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 721 | out: |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 722 | free(n); |
| 723 | free(rr); |
| 724 | free(rrtmp); |
Heiko Stuebner | eda753b | 2020-06-18 16:23:25 +0200 | [diff] [blame] | 725 | if (ret < 0) |
| 726 | rsa_free_key_prop(*prop); |
AKASHI Takahiro | e0d310b | 2020-02-21 15:12:58 +0900 | [diff] [blame] | 727 | return ret; |
| 728 | } |